MLLM-as-a-Judge — leaderboard
Evaluates multimodal LLMs as judges for scoring and ranking outputs, testing meta-evaluation capabilities across vision-language tasks.
Metric: Average Score. Source: mllm-judge.github.io. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | 0.607 | 0.9 |
| 2 | 0.804 | 0.9 |
| 3 | 0.696 | 0.8 |
| 4 | 0.493 | 0.6 |
| 5 | 0.657 | 0.6 |
| 6 | 0.454 | 0.5 |
| 7 | LLaVA-1.5-13b | 0.5 |
| 8 | LLaVA-1.5-13b | 0.5 |
| 9 | LLaVA-1.5-13b | 0.5 |
| 10 | 0.449 | 0.5 |
Interactive version: theaggregate.ai/benchmark?slug=mllm-as-a-judge · How the rankings work · Data refreshed daily, snapshot 2026-07-22.